There is no such methodology used in developing unbreakable codes, I guess.

There is no such methodology used in developing unbreakable codes, I guess.

Standard book on Research Methodology may be consulted.

Theoretically, the One Time Pad is unbreakable. Developed in 1882, no AI needed. Look more to Information Theory. Claude Shannon himself proved its robustness. The problem, though, is how you securely distribute the keys to trusted entities (avoiding attacks like Man-in-the-Middle and endpoint sniffing, etc.). This was the motive for the revolutionary Diffey-Hellman key exchange algorithm.

Thanks to George Slade for his answer.

The answer will give some direction for solving the problem.

To add to George Slade's answer, you would need a true random number generator (TRNG) to create the one time pads, in theory if you generate ciphertext XYZ using a one time pad based on a TRNG then for a brute force attacker all 3 letter words in the dictionary are equally probable in being the original plaintext.

The point raised by Jeremy Scerri may be applicable.